Soft X-ray microscopy with a cryo scanning transmission X...

Soft X-ray microscopy with a cryo scanning transmissionX-ray microscope: I. Instrumentation, imaging andspectroscopy

J. MASER* 1 , A. OSANNA*, Y. WANG*, C. JACOBSEN*, J. KIRZ*, S. SPECTOR* 2 , B. WINN* &

D. TENNANT²*Department of Physics and Astronomy, State University of New York at Stony Brook,

Stony Brook, New York 11794-3800, U.S.A.

²Lucent Technologies Bell Laboratories, Holmdel, New Jersey 07733-1988, U.S.A.

Key words. Cryo microscopy, frozen hydrated, soft X-ray microscopy,

spectromicroscopy, synchrotron radiation, zone plate.

Summary

We have developed a cryo scanning transmission X-ray

microscope which uses soft X-rays from the National

Synchrotron Light Source. The system is capable of imaging

frozen hydrated specimens with a thickness of up to 10 mm

at temperatures of around 100 K. We show images and

spectra from frozen hydrated eukaryotic cells, and a

demonstration that biological specimens do not suffer mass

loss or morphological changes at radiation doses up to about

1010 Gray. This makes possible studies where multiple images

of the same specimen area are needed, such as tomography

(Wang et al. (2000) Soft X-ray microscopy with a cryo

scanning transmission X-ray microscope: II. Tomography.

J. Microsc. 197, 80±93) or spectroscopic analysis.

Introduction

Soft X-rays have unique characteristics which make them

well suited for studies of organic specimens. Their short

wavelength (l < 2±5 nm) makes high resolution imaging

possible, and presently available X-ray optics produce the

®nest focus (20±50 nm) of electromagnetic radiation of any

wavelength (Schneider et al., 1995; Spector et al., 1997). By

operating in the `water window' spectral region between

the carbon K-edge at 284 eV and the oxygen K-edge at

540 eV, one can image samples in up to 10 mm of water or

ice, whereas electron microscopes require ice layers under

1 mm thick (Grim et al., 1998; Jacobsen et al., 1998) (see

Fig. 1). One also has access to spectroscopic signatures

which re¯ect the chemical bonding state of major low-Z

constituents, without the plural inelastic scattering back-

ground which is present in electron energy loss spectro-

scopy. The depth of ®eld of today's optics is well matched to

the typical dimensions of cell-sized specimens, allowing

for the acquisition of tomographic projections in the

microscope.

These characteristics have been known for some time

(Wolter, 1952; Sayre et al., 1997), and a large number of

laboratories are carrying out studies using X-ray micro-

scopes (see e.g. Kirz et al., 1995; Thieme et al., 1998).

However, as with other techniques using ionizing radiation

for probing the specimen, radiation damage limits studies of

hydrated specimens at room temperature (Foster et al.,

1992; Williams et al., 1993). Soft X-ray imaging at 50 nm

resolution typically involves a dose to the specimen of , 106

Gray. This dose must be compared with a dose level of

, 102 Gray where cell cultures die, , 104 Gray where

immediate changes are seen on initially living specimens

(Foster et al., 1992; Kirz et al., 1995), and 106ÿ107 Gray

where mass loss and shrinkage is seen even on chemically

®xed specimens (Williams et al., 1993). Radiation damage is

especially signi®cant for studies in which multiple images

must be taken of the same specimen, such as in tomography

(Haddad et al., 1994; Lehr, 1997; Wang et al., 2000) and

for spectroscopic mapping of chemical states (Ade et al.,

1992; Zhang et al., 1996). In addition, the dose required to

achieve suf®cient image contrast increases signi®cantly

with the resolution as more powerful X-ray optics become

available. To reduce structural changes caused by radiation

damage, most previous X-ray microscopy studies have been

done on dehydrated or chemically ®xed specimens. Even

though ®xation allows single images of organic specimens

to be made at moderate resolution levels, it causes changes

Journal of Microscopy, Vol. 197, Pt 1, January 2000, pp. 68±79.

Received 10 February 1999; accepted 30 May 1999

q 2000 The Royal Microscopical Society68

Correspondence to: Chris Jacobsen.

Present addresses: 1 Advanced Photon Source, Argonne National Laboratory,

Argonne, Illinois 60439, U.S.A and 2 Lucent Technologies Bell Laboratories,

Holmdel, New Jersey 07733-1988, U.S.A.

to the specimen morphology, and does not preserve lipid-

rich structures (Coetzee & van der Merwe, 1984; Stead et al.,

1992).

To overcome these limitations, we have built a cryo

scanning transmission X-ray microscope (cryo STXM)

(Maser et al., 1998). This system allows imaging and

spectroscopy of fully hydrated specimens at temperatures of

around 100 K. At these temperatures, specimens several

micrometres in thickness can be preserved in an amorphous

ice matrix. Structural damage caused by the ionizing

radiation is reduced owing to reduced quantum yield for

ionization of chemical bonds at these temperatures, and by

immobilizing free radicals which are created by ionizing

radiation, thus preventing secondary damage. This has

been exploited in electron microscopy for a long time (see

e.g. Dubochet et al., 1988), and ®rst was demonstrated in an

X-ray microscope by Schneider & Niemann (1994) and

Schneider (1998). Our ®rst measurements (see Fig. 7)

indicate a structural stability of the specimen for doses up to

1010 Gray, which represents a 104-fold increase in radiation

stability for hydrated biological specimens relative to ®xed

specimens at room temperature. This observation agrees

with X-ray microscopy results for a specimen in cold

nitrogen gas, as described by Schneider et al. (1995). This

dose level is at least an order of magnitude higher than the

one at which sample `bubbling' is observed in cryo electron

microscopy (Grimm et al., 1998)

The Stony Brook cryo STXM

The design of our cryo STXM draws considerably from the

experience of the cryo transmission electron microscope

(TEM) community. In cryo TEMs, one is able to mechani-

cally position cryo specimen holders at high precision and

with acceptable drift. The use of an airlock allows for rapid

specimen interchange, and good vacuum conditions (10ÿ6

to 10ÿ7 Torr) minimize specimen contamination buildup.

For our cryo STXM, we therefore chose to use a TEM airlock

(JEOL 1000 type) and cryo specimen holders (E.A. Fischione

and Oxford Instruments). However, a STXM has some

important differences from a TEM which have been

addressed in the following ways:

X Synchrotron X-ray beams propagate horizontally, rather

than vertically, and the working distance of soft X-ray zone

plates is small; this necessitates modi®cations to the usual

TEM cryo holder design.

X Focusing is accomplished by changing the distance from

the zone plate to the specimen, and an order sorting

aperture (OSA) must be placed between the zone plate and

the specimen to allow only X-rays in the desired focus to

reach the specimen.

X Moving the X-ray beam for scanning the specimen

involves mechanical motion of complex beamline compo-

nents or the X-ray optics, which poses signi®cant challenges

in the mechanical design of these components. Instead, we

chose to hold the optic ®xed and scan the specimen for

image acquisition. This requires motions of several milli-

metres for coarse positioning and large ®eld imaging at

micrometre resolution, and over a small range for high

resolution imaging.

We describe below how our cryo STXM has been designed

to meet these requirements. We also note that other

approaches have been used for cryo X-ray microscopy,

whereby the specimen is immersed in a cold nitrogen gas

environment (Schneider, 1998), but for our applications

which include nitrogen edge spectroscopy we preferred to

avoid the need to normalize signals carefully to correct for

absorption in such a cryogenic gas.

Layout

The cryo STXM operates at the X1A undulator beamline

(Winn et al., 1996) at the National Synchrotron Light

Source at Brookhaven National Laboratory (Fig. 2). The

beamline provides soft X-rays of high spectral brightness in

the energy range 250±700 eV with an energy resolution of

0´1±0´5 eV. The incident coherent X-rays are focused into a

small spot using a Fresnel zone plate, yielding a ¯ux of

2 ´ 107 photons sÿ1 in a diffraction limited spot. The

specimen is positioned into the focal spot, and raster

scanned for image acquisition. Placing the focusing optics

upstream of the specimen has the advantage of minimizing

q 2000 The Royal Microscopical Society, Journal of Microscopy, 197, 68±79

Fig. 1. X-ray/electron dose comparison. The required dose for ima-

ging 20 nm protein structures in different ice thicknesses is shown

for soft X-ray and 100 keV electron microscopy. Electrons are best

suited to observing thin structures (e.g. viruses in thin ice layers),

whereas X-rays offer exponentially increasing advantages for ice

thicknesses corresponding to whole cell dimensions. (Samples in

1±2 mm thick ice layers may become accessible to transmission

electron microscopy if 400 keV energy ®lters become available.)

For both X-rays and electrons, a detective quantum ef®ciency of

100% was assumed. The X-ray calculations are based on the

work of Sayre et al. (1977) and Schmahl & Rudolph (1987),

whereas the electron calculations of Jacobsen et al. (1998) used

Crewe & Groves (1974) and Langmore and Smith (1992).

SOFT X -RAY MICROSCOPY WITH A CRYO STXM 69

the radiation dose to the specimen. It also enables us to use

secondary X-ray induced processes such as X-ray induced

luminesence (Jacobsen et al., 1993; Moronne, 1999) as

contrast forming mechanisms in addition to absorption or

phase contrast. Spectra from small areas are obtained by

changing the X-ray wavelength and refocusing the zone

plate in a coordinated manner. The latter is necessary since

the focal length of a zone plate changes with the photon

energy.

For the development of 3D imaging, the cryo STXM has

been designed to allow a one-axis rotation of the specimen.

This allows us to record tilt series of specimens and obtain

tomographic reconstructions with high spatial resolution. A

demonstration of this capability on a frozen hydrated

specimen using the cryo STXM is described by Wang et al.

(2000).

Our cryo STXM operates at a pressure of 10ÿ6 to

10ÿ7 Torr. This has several advantages. First, no convective

heat exchange between the cooled specimen and the

environment takes place, and thermal drifts of the specimen

stage and nearby optics are minimized. Second, the full

spectral range of X-rays provided by the X1A beamline can

be exploited for imaging and spectroscopy, without addi-

tional X-ray absorption by a gaseous specimen environ-

ment. This results in a reduction of image noise, which is of

particular importance for spectroscopic applications. Third,

contamination of the cold specimen is kept at a minimum

when good vacuum conditions are maintained. In fact, we

were able to image the same specimen for time periods on

the order of 40 h without any noticeable build-up of

contamination.

The zone plate and the order sorting aperture (OSA) are

mounted inside the vacuum chamber (Fig. 4), where they

can be positioned with respect to the X-ray beam using ®ne

adjustment stages with micrometre replacement actuators

(New Focus piezomikes). The OSA is mounted at the end of

a thin (0´5 mm wide) strip, which is aligned parallel to the

rotation axis of the specimen. This allows us to rotate the

specimen around the OSA and achieve tilting angles of

6 608. The entire assembly of zone plate and OSA is

mounted on a linear stage inside the vacuum chamber to

allow focusing.

The vacuum chamber with the X-ray optics and specimen

stage is mounted on an optical table with vibration isolation

supports. The vibration isolation supports were specially

chosen to have better than 30 mm position reproducibility so

as to maintain stable alignment of the system to the 200±

300 mm wide X-ray beam from the beamline. The vacuum

chamber has a number of viewports to allow the microscope

operator to verify positioning of components. A turbomo-

lecular pump is used to pump the chamber to a pressure of

10ÿ6 Torr. To avoid mechanical vibrations and improve the

vacuum further, the turbo pump can be valved off, and the

system can be pumped with an ion getter pump which is

mounted directly to the chamber with an isolating gate

valve.

Focusing optics

Fresnel zone plates combine high spatial resolution, broad

wavelength coverage, ease of alignment and good stability.

Zone plates (ZPs) are circular diffraction gratings with

radially increasing line density (see e.g. Michette, 1986).

Incident X-rays are focused into a series of foci of positive

and negative diffraction orders. For imaging, the ®rst

positive diffraction order is usually chosen. Other diffraction

orders as well as undiffracted radiation contribute to

background signal in the image, and have to be blocked

from reaching the specimen by placing an OSA next to the

specimen (Fig. 2). The distance from the OSA to the

specimen is typically 30% of the focal length, and

determines the working distance.

The focusing ef®ciency of ZPs depends on the choice of

materials and on some geometrical factors (Kirz, 1974).

Typical soft X-ray ZPs are made of germanium or nickel,

and achieve a focusing ef®ciency on the order of 10±20%

(see e.g. Thieme et al., 1998). The smallest achievable spot

size is determined by the size drN of the outermost grating

lines. Using electron beam lithography and successive

pattern transfer, nickel ZPs with a thickness of 60 nm

with an outermost zone width of drN�18 nm and a

corresponding theoretical Rayleigh resolution of 22 nm

have been manufactured (Spector et al., 1997). Our highest-

resolution ZPs were previously fabricated with a maximum

outer diameter of d�80 mm, corresponding to a focal length

f � ddrN/l and working distance , f/3 on the order of

0´5 mm or less at an X-ray energy of 516 eV. However, the

Fig. 2. Optical set-up of the cryo STXM at the X1A beamline at the

NSLS. The X1A undulator delivers tuneable soft X-rays of high

spectral brightness from the storage ring. A monochromator

selects the desired energy and bandpass. A ZP is used to focus

the monochromatized beam to a small X-ray spot. An OSA blocks

background radiation not contained in the ®rst order zone plate

focus. The specimen is placed in the focal spot and raster scanned

for image acquisition. Focusing is achieved by moving ZP and OSA

along the X-ray beam direction.

70 J. M AS ER ET AL .


construction of the cryo specimen holder and the need for

tilting the specimen by angles in excess of 6 608 for

tomographic experiments required longer working distances

for cryo STXM. Therefore, we have developed ZPs with

diameters of 160 mm, and outermost zone widths drN of

60 nm and 45 nm with corresponding Rayleigh resolution

of 73 nm and 55 nm, respectively (Spector et al., 1998).

These provided us with working distances of a millimetre or

more at E > 500 eV, which considerably eased microscope

commissioning and initial tomographic experiments. We

believe that it will be possible to improve on this in the

future; for example, a ZP with 240 mm diameter and 45 nm

outermost zone width would have a usable working

distance of 0´8 mm when used at the carbon K edge.

Cryo specimen holder

Cryo specimens are prepared on electron microscope grids,

which are mounted in turn on the cryo specimen holder in

its workstation (E. A. Fischione Inc., Oxford Instruments).

The tip of the cryo holder with the specimen grid is cooled to

about 100 K by thermal conduction to liquid nitrogen in a

small dewar. The specimen temperature is continuously

monitored, and can be raised in a controlled way using a

built-in heater circuit. The heater can be used to freeze-dry

the specimen in-situ by slowly warming the specimen to

room temperature, or to investigate ice crystal formation in

the specimen by raising the temperature above the

recrystallization temperature of solid water. A shutter can

be placed over the specimen to protect it from contamina-

tion during transfer into the main vacuum chamber. The tip

of the specimen holder ends in a small ball of borosilicate

glass, which forms the interface between the sample and the

scanning stage. In comparison with a standard TEM cryo

holder, the dewar has been rotated by 908 to accommodate

the horizontal X-ray beam. The tip area has been stiffened

to better accommodate the forces caused by scanning the

specimen holder. The geometry of the tip has been modi®ed

to minimize interference with ZP and OSA during tilt, and to

reduce shadowing effects at large tilt angles.

Scanning mechanism

As mentioned previously, we scan the specimen through the

®xed X-ray focus to acquire images. In order to achieve

highest resolution, the mechanical precision and stability of

the scanning mechanism has to exceed the diffraction-

limited resolution provided by the X-ray optics. This requires

high precision in the positioning of the specimen, small

vibration amplitudes of all components relating to the


Fig. 3. Scanning mechanism of the cryo STXM as seen from the

direction of the incident X-ray beam. The cryo specimen holder is

inserted into the vacuum chamber through an airlock, which in

turn is connected to the vacuum chamber through bellows. A pre-

cision lever mechanism allows horizontal and vertical motion of

airlock and specimen holder. The tip of the cryo holder is placed

in a receptacle on the scanning stage, and held in place by an

adjustable preload. For high resolution scans, an in-vacuum ¯ex-

ure stage moves the specimen holder horizontally and vertically.

Coarse scans are performed by moving the whole ®ne stage and

the specimen holder using out-of-vacuum linear stages driven by

stepping motors.

Fig. 4. Side view of the cryo STXM. Monochromatized X-rays enter

the vacuum chamber from the left. The ZP and the OSA are

mounted on separate three-axis positioning stages for alignment.

Both stages are mounted to an in-vacuum stepping motor to

allow focusing. The coarse and ®ne scanning specimen stages are

placed on the upstream side of the vacuum chamber (see also

Fig. 3). The X-ray detector is mounted in a separate vacuum can

on a platform, which can be positioned using out-of-vacuum man-

ual positioners. Using an in-vacuum slide, the X-ray detector can

be replaced with a phosphor screen, a silicon photodiode, or an

in®nity-corrected visible light microscope objective. The micro-

scope head is located outside the vacuum chamber, along with

an alignment telescope. The vacuum chamber is supported on

kinematic mounts. This allows alignment of the chamber to the

X-ray beam without moving the scanning stages and the detector

set-up.


scanning mechanism or the X-ray optics, and small thermal

drifts between X-ray optics and specimen. Here we describe

the overall layout of the scanning mechanism and the

interface between specimen holder and vacuum chamber

(see Figs 3 and 4).

The scanning motions are provided by a combination of

coarse motion for covering large ®elds, and a ®ne stage for

high resolution. The coarse stage is formed by stepping

motors which are mounted outside the vacuum chamber,

and provide a motion of 12 mm ´ 9 mm with a resolution

of 0´1 mm. They are used to take overview scans of the

whole specimen grid, or to position into the X-ray beam

alignment features, which are mounted next to the

specimen on the cryo holder. For high-resolution scans,

we use a piezo-driven ¯exure stage with closed-loop

capacitance feedback, which covers a ®eld of 60 mm. Our

design is based on one used in an atomic force microscope,

which demonstrated 3 nm resolution and 1 : 104 linearity

in scans of a holographic grating (Lindaas, 1994; Lindaas

et al., 1996). In the cryo STXM, noticeable image

distortions are observed if the vacuum force of the airlock

bellows is not properly compensated for by springs

installed for this purpose (an area to be addressed in

future development of the instrument). This ¯exure stage is

positioned on a platform inside the vacuum chamber, and

this platform is coupled to the out-of-vacuum stepping

motors by three posts passing through welded bellows.

Coarse scans using the stepping motors move the airlock,

¯exure stage and cryo holder with specimen inside the

vacuum chamber.

The specimen holder is loaded into the cryo STXM using

a TEM-type side entry airlock (JEOL). The airlock is

connected to the vacuum chamber using welded bellows,

and suspended in a lever mechanism which allows vertical

and horizontal translation of the cryo holder tip about a

distant, well de®ned point. After transfer of the specimen

holder into the vacuum chamber, a capture nut provides a

tight connection between the cryo holder and the airlock,

so both are moved as one unit when moving and scanning

the specimen. This avoids any relative stick±slip motion

which would compromise the spatial resolution during

scanning. The ball at the tip of the specimen holder is

placed into a cone-shaped receptacle on the scanning

stage, thereby providing a precise kinematic link between

the two. The receptacle is made of invar, which, owing to

its low coef®cient of thermal expansion, reduces drifts

caused by the large temperature gradient between the cold

specimen holder and the scanning stage. A preload is

provided by adjustable springs on the airlock to keep the

specimen holder ®rmly placed in the receptacle on the

scanning stage during operation. The inner part of the

airlock can be rotated using a small actuator. This allows

rotation of the specimen holder inside the vacuum

chamber for acquisition of tomograpic data.

Detector and alignment optics

Our detector for transmitted X-rays is a screen coated with

P47 phosphor, and coupled through a light pipe to a

photomultiplier tube. A 100-nm thick silicon nitride

window coated on both sides with 40 nm of chromium is

placed at the entrance to the detector to make this system

relatively insensitive to visible light. The photomultiplier is

run in pulse counting mode, and water cooling pipes are

routed to its enclosure to minimize thermal noise in

operation. The overall ef®ciency of this system is only about

10%. We plan to replace this system with a more ef®cient

silicon drift detector and/or an avalanche photodiode in the

near future.

The X-ray detector is mounted to a platform inside the

vacuum chamber, which is connected through bellows-

sealed tripod posts to an out-of-vacuum manual translation

stage (Fig. 4). This allows positioning of the detector on the

X-ray axis. A sliding stage on top of the detector platform

can be externally manipulated to move the X-ray detector

aside and place other detectors in the beam. We use a large

area silicon photodiode in current mode for ¯ux monitoring,

and a phosphor screen for alignment purposes. Also, an

in®nity-corrected visible light microscope lens, which

operates in connection with a microscope head outside

the vacuum chamber, allows in-situ previewing of the cold

specimen. The microscope head and an alignment micro-

scope are installed at the downstream side of the vacuum

chamber. A mirror allows switching between the two

instruments. The telescope is used to view ZP, OSA and

specimen at adequate magni®cation for alignment pur-

poses. We achieve better than 100 mm pre-positioning of

these components, thereby signi®cantly speeding up the

®nal alignment step which uses X-rays.

Controls

The microscope is controlled by a Unix workstation. The

electronics and software system is similar to that used in a

scanning force microscope (Lindaas et al., 1996), where the

computer is involved in motion instruction and data

transfer once per scanline rather than at every pixel. Up

to eight channels of 16-bit analogue data can be collected,

and up to six channels of pulse rate data can be collected

with 16-bit dynamic range and rate prescaling. This allows

data collection from different detector channels (for example

from con®gured detectors), and can be used for recording

single measurements per pixel (for example temperature of

the sample). The graphical user interface is written in IDL

(Research Systems, Inc. Boulder, CO), which allows easy

control of the different operation modes of cryo STXM such

as 2D imaging, focus scans and acquisition of sequences of

energy-tuned images for spectromicroscopy (Jacobsen et al.,

2000), as well as display of ongoing data collection and

image processing.



Early operation and experimental results

Preparation of frozen hydrated biological specimens

Specimens are either grown or deposited on formvar-coated

gold grids. When cooling the specimen, the formation of ice

crystals has to be reduced as far as possible in order to avoid

damage to the morphology of the specimen. This can be

achieved by cooling at a high rate. We use plunge freezing, a

method where the specimen is plunged at high speed

(several m sÿ1) into a liquid cryogen (this freezing technique

is used by electron microscopists to prepare vitri®ed

specimens with little or no ice crystal artefacts; see e.g.

Dubochet et al. (1988)). For specimens several micrometres

in thickness, cooling rates on the order of 103ÿ104 K sÿ1

can be achieved, and ice crystal formation suppressed. As

cryogen, we use liquid ethane, which is cooled by liquid

nitrogen to a temperature of around 90 K, just above its

melting point. The use of liquid ethane greatly improves

specimen freezing, because gas bubbles would form around

the specimen and reduce thermal contact with the cryogen

if liquid nitrogen was used directly.

After plunging the grid with the specimen into the

cryogen, it is transferred to the cryo holder. This is done in a

workstation, where the specimen is kept under cold nitrogen

vapour above a liquid nitrogen reservoir to maintain low

specimen temperature and to avoid contamination. After

mounting, a small shutter on the cryo holder is closed over

the specimen for transfer into the airlock and successively

into the microscope's vacuum chamber. The shutter can

now be opened to begin operation.

Instrumental performance

We performed a number of tests to evaluate the perfor-

mance of the microscope. We tested the focusing properties

of the optics and the stability of the stage against vibrations

by imaging a resolution test pattern mounted on the

specimen holder. With a 160 mm diameter ZP with an outer

zone width of 45 nm we obtained the image shown in Fig. 5.

Periodic lines and spaces of 35 nm are visible, even though

the ZP was partially coherently illuminated. With a 45-nm

outermost zone width and an ideal object with 100%

contrast, one would expect to see periodic lines and spaces

as small as 23 nm if the ZP were fully coherently

illuminated. We conclude that the system is able to operate

near the X-ray optical resolution limit. We use data from

this and similar images to verify that the stage is linear,

orthogonal and correctly calibrated at the level required to

make quantitative measurements. We measured thermal

drifts by mounting a resolution test pattern on the specimen

holder and imaging it repeatedly at a temperature of 110 K.

From the apparent shift of the image as a function of time

we deduced an upper limit of the thermal drift of 4 AÊ sÿ1.

We veri®ed the thermal stability of the specimen by


Fig. 5. Images of a germanium test pattern demonstrate the ability of the cryo STXM to image high resolution structures. The test pattern

consists of 58 spokes, with rings located at radii where the spoke widths are 40, 60, 80, 140 nm, and so on as indicated. As can be seen at

left, the germanium test pattern is located within a 5-mm pinhole of nickel. The image at right was taken at 24 nm step size using a ZP with

drN �45 nm outermost zone width and diameter d�160 mm, showing spokes of width below 35 nm. The resolution was limited in this case

by partial spatial coherence of the X-ray beam.


monitoring the thermocouple built into the tip of the holder.

We looked for the build-up of contamination on the frozen

hydrated specimen by imaging the same area repeatedly

over a period of 40 h. We observed no change in mass

thickness over this period of time when the pressure in the

chamber was 0´5±2 ´ 10ÿ6 Torr.

Imaging of frozen hydrated 3T3 ®broblasts and study ofstructural damage of frozen hydrated specimens

EM grids containing live 3T3 ®broblast cells were taken

from the culture medium, brie¯y rinsed in phosphate-

buffered saline prior to plunge freezing, and transferred into

the microscope. The cells were imaged at a temperature

around 110 K. In practice, we end up with a layer of

vitreous ice that is not uniform. To ®nd areas where cells are

in ice of optimum thickness, we ®rst performed a coarse

exploratory scan that covered the entire grid. On this overall

map we recorded the coordinates of interest for higher

resolution imaging. We then recentred the image ®eld on

one of these coordinates. To focus, we collected a series of

repeated line scans near the cell of interest at decreasing

focal distances and selected the line with the sharpest

features (Wang et al., 2000). We then shifted to the cell for

high resolution imaging. Figure 6 shows an image of a

frozen hydrated 3T3 cell obtained in this way.

To study radiation damage to the specimen at cryogenic

temperature, we exposed a frozen hydrated ®broblast to

high radiation doses and looked for morphological changes

during and after exposure at low temperature, as well as

after warming the specimen to room temperature. After

initial imaging of a ®broblast, we increased the size of

various beamline apertures to greatly increase the X-ray

¯ux at a loss of spatial and spectral resolution. We let the

beam dwell on several specimen regions in and around the

nucleus for as long as 45 min per area for accumulated

Fig. 6. Image of an un®xed, frozen hydrated 3T3 ®broblast obtained in the early operation of our cryo STXM. The image is taken at an energy

of 516 eV, close to the oxygen absorption edge where transmission of water is high. The ®broblast was grown in culture medium on a form-

var-coated gold grid, brie¯y washed in a buffer solution, and plunge-frozen at high speed into liquid ethane at a temperature of 90 K. The

cell nucleus with several nucleoli can be clearly seen. Dense lipid vesicles can be seen surrounding the nucleus. The second image shows a

detail at the lower edge of the nucleus. Fine structure surrounding the carbon-dense lipid vesicles can be observed, and the nuclear mem-

brane can be seen. In the water surrounding the cell, as well as in the cytoplasm, groups of lamellar structures are visible. We believe these

are an indication of ice crystal formation in the layer of frozen water. Ice crystals could have formed if the water layer around the cell was

relatively thick prior to plunge-freezing. A typical plunge-frozen grid contains specimen areas of large water thickness which are virtually

opaque to X-rays, areas of moderate thickness where lamellar structures such as the above are observed, as well as areas where the

water layer is very thin and no artefacts are visible (see e.g. Fig. 8).



protein doses of more than 1010 Gray at a photon energy of

516 eV. Figure 7 shows an image of the ®broblast taken at

low temperature after radiation exposure to 1010 Gray. No

structural damage is obvious from this image, and no

structural changes are seen in comparison to an image

taken prior to the exposure. We then warmed the specimen

slowly to room temperature by heating the cryo holder. This

had the effect of freeze-drying of the specimen in the

vacuum chamber, allowing all specimen constituents of low

molecular weight to escape. The inset in Fig. 7 shows the

cell after warm-up, revealing the full extent of the radiation

damage inherent in the specimen areas that received high

doses. Most of the mass in these areas is lost, leading to a

series of holes in the cell structure. Even though radiation

damage on a molecular level was present in the cold

specimen after exposure to high radiation doses, it did not

manifest itself in mass loss or morphological changes at the

sub-100 nm resolution limit we worked at during these

measurements. Figure 8 shows another cell which was

selectively exposed to a dose of around 5 ´ 1011 Gray. At this

dose level, the exposed area shows slight mass loss even

without warming the specimen to room temperature,

indicating a limit of structural stability of frozen hydrated

specimens in a cryo STXM at doses exceeding 1011 Gray. We

note that we have yet to see signs of specimen bubbling even

at these selectively high doses, whereas in cyro electron

microscopy such bubbling is seen at electron exposures of

, 104 eÿ nmÿ2 at 100 keV or about 4 ´ 108 Gray. It is unclear

whether this might be due to the lower dose rate delivered

here, to the lower ratio of ice to organic absorption with

`water window' soft X-ray illumination, or some other reason.

Spectromicroscopy

Soft X-rays are much more likely to be absorbed than

elastically scattered, and the cross-section for inelastic


Fig. 7. Demonstration of the radiation dose tolerance of a frozen hydrated 3T3 ®broblast which was rapidly plunged into liquid ethane. Fol-

lowing acquisition of a 100-nm resolution image at a temperature of 110 K, an intense, , 0´2 mm beam spot was left stationary on several

locations (noted with arrows) for several minutes (in particular, several spots on the nucleolus were given a dose of about 1010 Gray, in con-

trast to the , 106 Gray delivered to the cell over the course of many images). The image at left was acquired after these exposures, with no

discernible difference from an initial image which is not shown here. The specimen was then allowed to slowly warm up to room tempera-

ture. The smaller insert shows an image of the cell taken after warm-up. The heavily dosed regions lost signi®cant mass, presumably due to

the reaction of radiation-produced free radicals as the temperature approached room temperature, while the rest of the cell shows less pro-

nounced artefacts from radiation damage and freeze-drying. No particular freeze-drying protocol was followed when warming the specimen,

which may have led to artefacts in the image.


scattering is negligibly small by comparison. As a result,

images are largely free from multiple scattering effects and

provide easily quanti®able maps of organic mass. Further-

more, one can exploit X-ray absorption near edge spectro-

scopy (XANES) resonances, which are caused by excitations

of inner shell electrons into available molecular orbital

states, to obtain selective contrast of certain chemical states

of atoms which are present at high concentration. The

combination of high resolution imaging with this sensitive

spectral information is termed XANES spectromicroscopy. In

the soft X-ray range, this approach has been exploited in

applications including the characterization of polymers

(Ade et al., 1992) and the mapping of protein/DNA

concentrations in biological specimens (Zhang et al.,

1996), in addition to other applications cited in a recent

review (Ade, 1998). Because there is no background of

plural inelastic scattering, the signal-to-noise ratio is much

higher than in electron energy loss spectroscopy (Isaacson

& Utlaut, 1978; Rightor et al., 1997).

The natural energy width of soft X-ray XANES reso-

nances is about 0´3 eV, so for quantitative studies one

requires correspondingly high energy resolution. This

requirement is met using a spherical grating monochro-

mator at the X1A beamline (Winn et al., 1996). Studies to

date have been conducted mostly at the carbon absorption

edge. Work at the nitrogen and oxygen edges has been

dif®cult owing to the residual air present in the helium-

¯ushed sample environment in our room temperature

STXM (Kirz et al., 1995). With the specimen placed in the

vacuum of our cryo STXM, we are now able to pursue

spectromicroscopy through the whole spectral range

provided by the X1A beamline.

For biological applications, it may be possible to predict

the dominant features of the near-edge absorption spectra of

proteins from previously acquired amino acid absorption

spectra (Kirtley et al., 1992; Boese et al., 1997), and

therefore estimate from knowledge of amino acid sequences

the likely contrast for chemical state mapping experiments

of proteins. These applications require that several images

be taken of the same object at different photon energies, so

that the radiation dose stability of the cryo method is needed

for high resolution studies of hydrated specimens. In Fig. 9,

Fig. 8. An example of the effects of a dose in excess of 1011 Gray on a frozen hydrated specimen. The insets show a detail at the right edge of

the nucleus. The image in the ®rst inset is taken at the resolution limit of the 60 nm ZP (total accumulated dose roughly 106 Gray). A small

area in the centre was exposed to an accumulated dose of approximately 5 ´ 1011 Gray after this image was taken. The second inset shows the

same detail as before, with the specimen still at cryogenic temperature. An area of reduced density can now be seen in the area where the

focused X-ray beam dwelled, which we attribute to structural damage of the cryogenic specimen. This is an indication of the dose level at

which morphological changes will be encountered in frozen hydrated specimens in a cryo STXM.



we show the absorption spectrum of glutamic acid versus

that of ice taken in cryo STXM. It illustrates that the

resonance near 531 eV due to the C�O bond in glutamic

acid is well separated in energy from the oxygen edge in ice.

This should make it possible to perform XANES imaging of

the organic components of hydrated specimens at the

oxygen absorption edge in the presence of several micro-

metres of ice. Studies of this type are now in progress. An

important consideration for these studies will be the

radiation dose sensitivity of near-edge absorption reso-

nances in organic materials at cryo temperatures. It is

known that ®lms of polymethyl methacrylate at room

temperature suffer a 15% decline in C�O bond strength

following absorbed doses of 107 Gray at 290 eV (Zhang et al.,

1995), but this polymer is especially sensitive to radiation

exposure, which is why it is commonly used as a

photoresist. We are now beginning studies of the radiation

dose sensitivity of XANES resonances at liquid nitrogen

temperatures.

Discussion

We have presented early results from operation of the cryo

STXM. The system has unique characteristics which are

advantageous for studies where 2D and 3D imaging of

labelled or unlabelled structures is desired of un®xed,

hydrated specimens with the 1±10 mm thickness typical of

eukaryotic cells. In addition, resonances in absorption

spectra can be used for mapping of chemical states of low-

Z atoms at high spatial resolution.

Images of unmodi®ed samples of the sort shown in Figs 6±

8 are dif®cult to acquire by other means. Optical microscopy

of unlabelled regions of cells tends to have a point-to-point

resolution $ 200 nm; , 100 nm resolution can only be

obtained using interferometric and deconvolution techni-

ques on ¯uorescently labelled structures (Carrington et al.,

1995; Hell et al., 1997). Near-®eld (Betzig et al., 1991) and

related (Zenhausern et al., 1995) microscopes deliver higher

resolution images, but only of near-surface features. Using

1 MeV (O'Toole et al., 1993) or energy ®ltered 120 keV

(Grimm et al., 1998) electron microscopes, one can study

samples at higher resolution in ice as thick as 0´5±1 mm,

but this thickness limit is not suf®cient for most intact

eukaryotic cells. Cryo X-ray microscopy offers a way to

study the 2D and 3D organization of larger scale aggregates

in unmodi®ed frozen specimens at a resolution intermediate

between the capabilities of optical and electron microscopes.

In addition, efforts are underway to develop labelling

methods appropriate for X-ray microscopes (Jacobsen et al.,

1993; Moronne et al., 1994; Chapman et al., 1996;

Moronne, 1999).

Our ®rst studies with the cryo STXM were taken using

ZPs of lower resolution and correspondingly larger focal

length than the present state of the art. We expect to

improve the spatial resolution as methods for fabrication of

ZPs with large diameters advance, and as additional

operational experience allows us to use high resolution

ZPs with shorter focal length. We also expect to improve the

ef®ciency and the dynamic range of our X-ray detector to

speed data acquisition and improve image statistics. We

®nally plan to adopt other contrast mechanisms such as X-

ray-induced luminescence and phase contrast methods for

use in our cryo STXM. The experimental program pursued

with this system will involve signi®cant efforts in micro-

spectroscopy, spectromicroscopy and nanotomography, and

could be extended to specimens such as polymers and other

radiation-sensitive soft matter.

Acknowledgements

We thank Raymond Fliller, Konstantin Kaznacheyev, Jan

Warnking, Matthias Weigel and Sue Wirick for their

contributions to cryo STXM, Azeddine Ibrahimi for his

help with cell culture, and Marc Adrian, Wolfgang


Fig. 9. Soft X-ray absorption spectroscopy provides information on

the chemical state of organic compounds. The absorption spectrum

of , 100 nm thick ice (which we believe is in the amorphous

rather than the crystalline state) is shown, along with the absorp-

tion spectrum of a , 200 nm thick ®lm of glutamic acid. Both spec-

tra were normalized to tabulated absorption coef®cients Henke et al.

(1993) at 510 and 570 eV using a procedure described for example

by Boese et al. (1997), and assumptions of thin ®lm densities of

0´95 g cmÿ3 and 1´35 g cmÿ3 for ice and glutamic acid, respec-

tively. The absorption resonance at , 531 eV is thought to be

caused by C�O bonds in the amino acid, and the exact energy at

which it appears may differ among amino acids. Absorption edge

spectra of amino acids can be used to predict the absorption spectra

of organic complexes Boese et al. (1997), and images acquired

using near-edge absorption resonances can be used to form maps

of chemical bonding states (Ade et al., 1992), including quantita-

tive maps of protein and DNA distribution (Zhang et al., 1996).


Baumeister, Frank Booy, Jacques Dubochet, Robert Glaeser,

Rainer Hegerl, Richard Henderson, Richard Leapman,

Jonathan Sedat and Dieter Typke for many helpful

discussions. We gratefully acknowledge support from the

Of®ce of Biological and Environmental Research, US DoE

under contract DE-FG02-89ER60858, the National Science

Foundation under grants DBI-9605045 and ECS-9510499,

and the Alexander von Humboldt Foundation (Feodor-

Lynen Fellowship, JM). This work was carried out at the

National Synchrotron Light Source at Brookhaven National

Laboratory, which is supported by the US Department of

Energy.

References

Ade, H. (1998) X-ray spectromicroscopy. Experimental Methods in

the Physical Sciences, Vol 32 (ed. by R. Celotta and T. Lucatorto),

pp. 225±262. Academic Press, New York.

Ade, H., Zhang, X., Cameron, S., Costello, C., Kirz, J. & Williams, S.

(1992) Chemical contrast in X-ray microscopy and spatially

resolved XANES spectroscopy of organic specimens. Science, 258,

972±975.

Betzig, E., Trautman, J.K., Harris, T.D., Weiner, J.S. & Kostelak, R.L.

(1991) Breaking the diffraction barrier: optical microscopy on a

nanometric scale. Science, 251, 1468±1470.

Boese, J., Osanna, A., Jacobsen, C. & Kirz, J. (1997) Carbon edge

XANES spectroscopy of amino acids and peptides. J. Electron

Spectrosc. Relat. Phenom. 85, 9±15.

Carrington, W.A., Lynch, R.M., Moore, E.D.W., Isenberg, G.,

Fogarty, K.E. & Fay, F.S. (1995) Superresolution three-dimen-

sional images of ¯uorescence in cells with minimal light

exposure. Science, 268, 1483±1487.

Chapman, H.N., Jacobsen, C. & Williams, S. (1996) A character-

isation of dark-®eld imaging of colloidal gold labels in a scanning

transmission X-ray microscope. Ultramicroscopy, 62 (3), 191±

213.

Coetzee, J. & van der Merwe, C.F. (1984) Extraction of substances

during glutaraldehyde ®xation of plant cells. J. Microsc. 135,

147±158.

Crewe, A.V. & Groves, T. (1974) Thick specimens in the CEM and

STEM. I: Contrast. J. Appl. Phys. 45, 3662±3672.

Dubochet, J., Adrian, M., Chang, J.J., Homo, J.-C., Lepault, J.,

McDowell, A.W. & Schultz, P. (1988) Cryo-electron microscopy of

vitri®ed specimens. Q. Rev. Biophys. 21, 129±228.

Foster, G.F., Buckley, C.J., Bennett, P.M. & Burge, R.E. (1992)

Investigation of radiation damage to biological specimens at

water window wavelengths. Rev. Sci. Instrum. 63, 599±600.

Grimm, R., Singh, H., Rachel, R., Typke, D., Zillig, W. & Baumeister,

W. (1998) Electron tomography of ice-embedded prokaryotic

cells. Biophys. J. 74, 1031±1042.

Haddad, W.S., McNulty, I., Trebes, J.E., Anderson, E.H., Levesque,

R.A. & Yang, L. (1994) Ultra high resolution X-ray tomography.

Science, 266, 1213±1215.

Hell, S.W., Schrader, M. & van der Voort, H.T.M. (1997) Far-®eld

¯uorescence microscopy with three dimensional resolution in

the 100-nm range. J. Microsc. 187, 1±7.

Henke, B.L., Gullikson, E.M. & Davis, J.C. (1993) X±ray interac-

tions: Photoabsorption, scattering, transmission and re¯ection at

E�50±30 000 eV, Z�1±92. Atomic Data Nucl. Data Tables, 54,

181±342.

Isaacson, M. & Utlaut, M. (1978) A comparison of electron and

photon beams for determining micro-chemical environment.

Optik, 50, 213±234.

Jacobsen, C., Flynn, G., Wirick, S. & Zimba, C. (2000) Soft X-ray

spectroscopy from sub-100 nm regions. J. Microsc. in press.

Jacobsen, C., Lindaas, S., Williams, S. & Zhang, X. (1993) Scanning

luminescence X-ray microscopy: imaging ¯uorescence dyes at

suboptical resolution. J. Microsc. 172, 121±129.

Jacobsen, C., Medenwaldt, R. & Williams, S. (1998) A perspective

on biological X-ray and electron microscopy. X-ray Microscopy

and Spectromicroscopy (ed. by J. Thieme, G. Schmahl, E. Umbach

and D. Rudolph), pp. II±93±102. Springer-Verlag, Berlin.

Kirtley, S.M., Mullins, O.C., Chen, J., van Elp, J., George, S.J., Chen,

C.T., O'Halloran, T. & Cramer, S.P. (1992) Nitrogen chemical

structure in DNA and related molecules by X-ray absorption

spectroscopy. Biochim. Biophys. Acta, 1132, 249±254.

Kirz, J. (1974) Phase zone plates for X-rays and the extreme UV. J.

Opt. Soc. Am. 64, 301±309.

Kirz, J., Jacobsen, C. & Howells, M. (1995) Soft X-ray microscopes

and their biological applications. Q. Rev. Biophys. 28 (1), 33±

130. Also available as Lawrence Berkeley Laboratory report

LBL±36371.

Langmore, J.P. & Smith, M.F. (1992) Quantitative energy-®ltered

electron microscopy of biological molecules in ice. Ultramicro-

scopy, 46, 349±373.

Lehr, J. (1997) 3D X-ray microscopy: tomographic imaging of

mineral sheaths of bacteria Leptothrix ochracea with the

GoÈttingen X-ray microscope at BESSY. Optik, 104 (4), 166±170.

Lindaas, S.A. (1994) X-ray Gabor holography using a scanning force

microscope. PhD Thesis, Department of Physics, State University

of New York at Stony Brook.

Lindaas, S., Howells, M., Jacobsen, C. & Kalinovsky, A. (1996) X-

ray holographic microscopy by means of photoresist recording

and atomic-force microscope readout. J. Opt Soc. Am., A, 13 (9),

1788±1800.

Maser, J., Jacobsen, C., Kirz, J., Osanna, A., Spector, S., Wang, S. &

Warnking, J. (1998) Development of a cryo scanning x-ray

microscope at the NSLS. X-ray Microscopy and Spectromicroscopy

(ed. by J. Thieme, G. Schmahl, E. Umbach and D. Rudolph), pp.

I±35±44. Springer-Verlag, Berlin.

Michette, A.G. (1986) Optical Systems for Soft X-Rays. Plenum, New

York.

Moronne, M.M. (1999) Development of X-ray excitable lumines-

cent probes for scanning X-ray microscopy. Ultramicroscopy, 77,

23±36.

Moronne, M.M., Larabell, C., Selvin, P.R. & Irtel von Brenndorff, A.

(1994) Development of ¯uroescent probes for X-ray microscopy.

Proceedings of the 52nd Annual Meeting of the Microscopy Society of

America (ed. by G. W. Bailey and A. J. Garratt-Reed), pp. 48±49.

San Francisco Press, San Francisco.

O'Toole, E., Wray, G., Kremer, J. & McIntosh, J.R. (1993) High

voltage cryomicroscopy of human blood platelets. J. Struct. Biol.

110, 55±66.

Rightor, E.G., Hitchcock, A.P., Ade, H., Leapman, R.D., Urquhart,



S.G., Smith, A.P., Mitchell, G., Fischer, D., Shin, H.J. & Warwick,

T. (1997) Spectromicroscopy of poly(ethylene terephthalate):

comparison of spectra and radiation damage rates in X-ray

absorption and electron energy loss. J. Phys. Chem., B, 101 (11),

1950±1960.

Sayre, D., Kirz, J., Feder, R., Kim, D.M. & Spiller, E. (1977) Potential

operating region for ultrasoft X-ray microscopy of biological

specimens. Science, 196, 1339±1340.

Schmahl, G. & Rudolph, D. (1987) Proposal for a phase contrast X-

ray microscope. X-Ray Microscopy: Instrumentation and Biological

Applications (ed. by P. C. Cheng and G. J. January), pp. 231±238.

Springer-Verlag, Berlin.

Schneider, G. (1998) Cryo X-ray microscopy with high spatial

resolution in amplitude and phase contrast. Ultramicroscopy, 75,

85±104.

Schneider, G. & Niemann, B. (1994) Cryo X-ray microscopy: ®rst

images of specimens at low temperatures. X-Ray Sci. 2, 8±9.

Summer 1994 newsletter, Centre for X±ray Science, King's

College, London.

Schneider, G., Niemann, B., Guttmann, P., Rudolph, D. & Schmahl,

G. (1995) Cryo X-ray microscopy. Synchrotron Radiation News, 8

(3), 19±28.

Schneider, G., Schliebe, T. & Aschoff, H. (1995) Cross-linked

polymers for nanofabrication of high-resolution zone plates in

nickel and germanium. J. Vacuum Sci. Technol., B, 13 (6), 2809±

2812.

Spector, S., Jacobsen, C. & Tennant, D. (1997) Process optimization

for production of sub-20 nm soft X-ray zone plates. J. Vacuum Sci.

Technol., B, 15 (6), 2872±2876.

Spector, S.J., Jacobsen, C.J. & Tennant, D.M. (1998) Zone plates for

a scanning transmission x-ray microscope. X-ray Microscopy and

Spectromicroscopy (ed. by J. Thieme, G. Schmahl, E. Umbach and

D. Rudolph), pp. IV±13±19. Springer-Verlag.

Stead, A.D., Cotton, R.A., Page, A.M., Dooley, M.D. & Ford, T.W.

(1992) Visualization of the effects of electron microscopy

®xatives on the structure of hydrated epidermal hairs of tomato

(Lycopersicum peruvianum) as revealed by soft X-ray contact

microscopy. Soft X-ray Microscopy, Vol 1741 (ed. by C. Jacobsen

and J. Trebes), pp. 351±362. Society of Photo-Optical Instru-

mentation Engineers (SPIE), Bellingham, Washington.

Thieme, J., Schmahl, G., Umbach, E. & Rudolph, D. eds. (1998) X-

Ray Microscopy and Spectromicroscopy. Springer-Verlag, Berlin.

Wang, Y., Jacobsen, C., Maser, J. & Osanna, A. (2000) Soft X-ray

microscopy with a cryo STXM: II. Tomography. J. Microsc. 197,

80±93.

Williams, S., Zhang, X., Jacobsen, C., Kirz, J., Lindaas, S., van'tHof,

J. & Lamm, S.S. (1993) Measurements of wet metaphase

chromosomes in the scanning transmission X-ray microscope.

J. Microsc. 170, 155±165.

Winn, B., Ade, H., Buckley, C., Howells, M., Hulbert, S., Jacobsen,

C., Kirz, J., McNulty, I., Miao, J., Oversluizen, T., Pogorelsky, I. &

Wirick, S. (1996) X1A: second generation undulator beamlines

serving soft X-ray spectromicroscopy experiments at the NSLS.

Rev. Sci. Instrum. 67 (9), 1±4 Paper A31.

Wolter, H. (1952) Spiegelsysteme streifenden Einfalls als abbildende

Optiken fuÈ r RoÈntgenstrahlen. Ann. Phys. 10, 94±114.

Zenhausern, F., Martin, Y. & Wickramasinghe, H.K. (1995)

Scanning interferometric apertureless microscopy: optical ima-

ging at 10 Angstrom resolution. Science, 269, 1083±1085.

Zhang, X., Balhorn, R., Mazrimas, J. & Kirz, J. (1996) Mapping and

measuring DNA to protein ratios in mammalian sperm head by

XANES imaging. J. Struct. Biol. 116, 335±344.

Zhang, X., Jacobsen, C., Lindaas, S. & Williams, S. (1995) Exposure

strategies for PMMA from in situ XANES spectroscopy. J. Vacuum

Sci. Technol., B, 13 (4), 1477±1483.



Soft X-ray microscopy with a cryo scanning transmission X...

Documents

Transcript of Soft X-ray microscopy with a cryo scanning transmission X...